遗传算法与模拟退火算法在神经网络.pdfVIP

第 30 卷 第 6 期 东 　北 　林 　业 　大 　学 　学 　报 Vol. 30 No. 6 　 2002 年 11 月 JOURNAL OF NORTHEAST FORESTRY UNIVERSITY Nov. 2002 ) 遗传算法与模拟退火算法在神经网络优化中的性能分析1 曹　军　苏建民　孙丽平　胡昆仑 ( 东北林业大学 ,哈尔滨 ,150040) 　　摘 　要 　神经网络有以任意精度逼近未知函数的能力 ,所以被广泛应用于各种领域中。目前广泛应用于神经 ( ) 网络优化的方法是反向传播 Back Propagation , BP ,但是 BP 的全局搜索能力很有限 ,而全局搜索方法是神经网络 ( ) ( 优化问题很有潜力的办法。文中研究了两种全局优化算法 :遗传算法 Genetic Algorithm , GA 和模拟退火 Simulated Annealing ,SA) ,并且比较了它们在神经网络优化中的性能。 关键词 　神经网络 ;优化 ;遗传算法 ;模拟退火 ;全局搜索 分类号 　O224 Performation Analysis on Optimization of Neural Network by Genetic Algorithm and Simulated Annealing/Cao Jun , Su ( ) Jianmin , Sun Liping , Hu Kunlun Northeast Forestry University , Harbin 150040 , P. R. China / / Journal of Northeast Forestry ( ) University. - 2002 , 30 6 . - 26～28 The neural network has the ability to closely approximate unknown functions to any degree of desired accuracy , so it apply to ( ) various kinds of fields extensively. Today , Back - propagation BP is the most widely used optimization techniques for training neural networks , it has been shown that BP severely limited in their ability to find global solutions , global search techniques have been identified as a potential solution to this problem. In this paper , the authors examine two global search techniques : Ge